We are bittensor we are coming!! First they ignore you.... Than they laugh at you... They fight you.... WE WIN!! #bittensor $tao @opentensor @const_reborn

Here's a draft of the tech report on the model training method I've been experimenting with, "Parallax". chutes.ai/parallax.pdf TL;DR: MoE models' params are mostly routed experts, and you can massively reduce VRAM and FLOPS per participant by splitting up those experts. You can also offload the expert training to commodity hardware further saving compute/VRAM per island. The crazy cool thing about these sketches is, you can actually onboard workers nearly instantly (sync time with ternary weights is a few MB), and they never need to download or stream the raw datasets (sketches contain all the work they need to do and are tiny). You'd probably want the first couple layers of the model in your own infra if you had sensitive data because otherwise you could do gradient inversion attacks to reconstitute the raw text, but beyond the first couple layers and not knowing which layer/expert you're training I think it's infeasible so privacy is pretty baked in. Decoupled DiLoCo/RDA-diloco style backbone sync, surrogates for non-owned routed experts with low rank updates to sync those, tiered sync cadences for various components, "sketches" to offload expert work, etc. 20b tested two different ways, plenty of small model iterations, and 176b params just to prove out feasibility. There are hundreds of additional experiments and loads of data we could also highlight as well, but the guts are there. Variations: - freeze routed expert weights instead of using surrogates, eliminates adam state, backward pass, etc., though you'd need different sync methods vs. low rank updates to the surrogates (the surrogates are already tiny, rank-8 updates to those even smaller) - hierarchical parallax, i.e. each node itself becomes an rda diloco style multi-learner which then syncs with the outer/global islands (the point here is to enable GPUs without NVLink/etc. and reduce GPU<=>GPU comms to maximize MFU on less-capable/commodity GPUs) - pipeline parallelize each island itself such that each island can decompose the backbone etc.

We are being recognized 🙏 hard work pays off. ❤️ $tao #bittensor

one word $tao

Bittensor Daily Recap - Jun 13, 2026 TAO: $266.67 ( 23.0%) Top gainer: Liquidity 21.5% Top loser: Bitstarter #1 -17.4% 128 active subnets tracked on subnetradar.com #bittensor $TAO #dTAO

$TAO daily snapshot. Today's movers, top chain buys, and where teams are shipping. Free on subnetradar.com #bittensor #dTAO

while SN77 Liquidity ripped 21.1%, SN110 Green Compute #bittensor #dTAO

Sunrise Digest · 2026-06-13 Liquidity Surges, Smart Money Moves, Risk Radar Blinks • SN77 Liquidity soars 19.1%, leading 24h gains with 335τ chain buys • Top wallets exit SN0, SN24 Quasar, SN64 Chutes, and SN3 Training, signaling smart money movement • Risk radar flags SN40 Chunking (65/100), SN69 Subnet 69 (55/100), SN15 ORO (51/100), and SN84 Subnet 84 (49/100) subnetradar.com/alpha/report… #bittensor $TAO #dTAO

The US just forced Anthropic to kill their newest models. @const_reborn predicted this outcome years ago. You can't ban Bittensor, you can't stop decentralized permissionless intelligence. Shared by everyone, owned by no one. #bittensor $TAO @opentensor

Every Bittensor subnet's alpha token halves its emission at 10.5M minted. Almost nobody is tracking it. The TAO base halving already happened in December 2025. The per-subnet alpha halvings are still ahead, and each one is independent: a subnet halves when its own cumulative alpha crosses 10.5M, on its own pace. When it crosses, the alpha it pays out per block drops by half. Where it stands across all 128 subnets right now: none have halved yet, and the pack sits roughly halfway there. Median 44% minted, 21 subnets already past the 50% mark. On current emission rates the first halvings are estimated around late 2027, but that is a projection, not a fixed date. So we built the radar. It ranks every subnet by estimated time to its next halving, shows the wave by quarter, and lets you paste a coldkey to see which of your own positions halve first. Verified against the subtensor source, real on-chain supply only. See where you stand: subnetradar.com/alpha/halvin… $TAO #bittensor #dtao

Excited to announce a new partnership between @NameTensor and @SubnetRadarcom 🤝 SubnetRadar provides real-time analytics, health monitoring, and risk assessment for subnets across the #bittensor ecosystem. Through this integration, you can now: 🟣 Search wallets using .tao names 🟣 View NameTensor badges across the platform 🟣 Track portfolios and Smart Money wallets by identity 🟣 Send $TAO directly to supported identities via @tip_tao_ Bringing identity directly into the tools people use every day, making #bittensor more accessible, discoverable, and user-friendly 🪪⚡️ 🔗 subnetradar.com

This is why we need decentralised AI like Bittensor. What a joke.

We rebuilt SubnetRadar from the ground up. New radar deck, full subnet dossiers, halving countdowns, conviction tracking, child hotkeys nobody else covers, and a Telegram bot that sends clean cards instead of noise. Everything is free. Full tour below. #bittensor $TAO #dTAO

Our telegram channels is be given a make over. For excisting users please typ / start for new menu. You pick your language by tapping, not by guessing. English, Chinese, French, Russian and Spanish. t.me/SubnetRadarBot $TAO

gm. you've been asking what we're building. time to pull back the curtain a little. the roadmap is public - your prompts are not. saygm.com/#roadmap

Bittensor Daily Recap - Jun 12, 2026 TAO: $213.64 ( 2.9%) Top gainer: Plaτform 20.5% Top loser: EvolAI -12.3% 128 active subnets tracked on subnetradar.com #bittensor $TAO #dTAO

Sunrise Digest · 2026-06-12 Green Compute Surge & EvolAI Slump; Smart Money Flows • Green Compute subnet (SN110) surged by 13.8% 24h, leading the ecosystem, potentially signaling increased interest in eco-friendly computing initiatives. • EvolAI subnet (SN47) experienced a significant drop of -14.4% 24h, indicating a potential shift in sentiment or market dynamics within the AI subnet. • Smart money clusters showed top wallets exiting SN0 and SN27, and accumulating in SN24 and SN95, suggesting strategic positioning by influential market participants. subnetradar.com/alpha/report… #bittensor $TAO #dTAO

$TAO daily snapshot. What moved · chain buys · dev signals — at a glance. Live on subnetradar.com #bittensor #dTAO

While SN100 Plaτform ripped 19.4% (pool 3881τ), SN47 EvolAI bled 11.8% (pool 1571τ). #bittensor #dTAO